Estimation of transition probabilities for diabetic patients using hidden Markov model

被引:0
|
作者
Manoj Kumar Varshney
Ankita Sharma
Komal Goel
Vajala Ravi
Gurprit Grover
机构
[1] University of Delhi,Department of Statistics, Hindu College
[2] University of Delhi,Department of Statistics, Faculty of Mathematical Sciences
[3] University of Delhi,Department of Statistics, Lady Shri Ram College
来源
International Journal of System Assurance Engineering and Management | 2020年 / 11卷
关键词
Diabetes; HbA1c; Hidden Markov model (HMM); Emission distribution; Goodness of fit;
D O I
暂无
中图分类号
学科分类号
摘要
Diabetes is a common non-communicable disease affecting substantial proportion of adult population. This is true, especially in developing countries like India thereby posing a huge economic burden not only on the patient’s family but also on the nation as a whole. In this paper, we have employed a hidden Markov model to estimate the transition probabilities between three states of diabetes and applied it to real life data. A total of 184 Type 2 diabetic patients were included in this study. These patients are classified in different states on the basis of their available baseline value of Hemoglobin A1c (HbA1c). A HMM fits well to the data by capturing the misclassified states, and shows that the patients who had HbA1c ≥ 6.5% have minimum chance of recovery and substantially higher risk of complications. All the statistical analysis has been performed using the “Hidden Markov” package in R software.
引用
收藏
页码:329 / 334
页数:5
相关论文
共 50 条